Hello!
I think I understand the sampling flaw you are getting at
of those who were ill, 60% were found to be watching cat videos
but general population, only 10% watch them. So we might conclude that watching cat videos causes illness
but what if amongst general population, 60% have watched the cat videos
so we might say there is no correlation
if I apply this same logic to option (D),
if the proportion of normal drivers with prosinopsis is just 2%
whereas the ones involved in accident is 4%
therefore, for the disease to cause the accident, it is assumed that these drivers have to be diagnosed with the disease?
If you please just could elaborate a little further,
this sampling flaw will be absolutely clear to me
Thank you so much!!
DmitryFarber
It's an odd scenario, but what they're getting at is a sampling flaw. Apparently, we know for a fact that 4% of those who had an accident had this condition. But for the general population, we just know that 2% have been diagnosed. Maybe the issue is that when someone has an accident, we check for this condition. Maybe 4% or more of drivers do have it, but we don't find out until they get in an accident. In fact, the condition may be quite common and may not be affecting the likelihood of an accident at all.
Let's look at an analogous case. What if I have a theory that seeing cat videos cause illness? So when someone gets sick, I go through their browsing history for the prior week and--lo and behold--60% of them have watched a cat video. But when I ask random people if they've watched a cat video in the past week, only 10% say yes. Aha! Cat videos show up disproportionately among people who got sick, so they must be causing a problem, right? But wait! The groups aren't the same. The larger group was just self-reporting, whereas I was actually *checking* those who got sick. Maybe if I checked *everyone's* phone, I'd find that many people had actually watched a cat video, even if they didn't remember. I might find that this happened 60% of the time, meaning there was no correlation between the videos and illness. Or maybe I'd find cat videos 95% of the time, meaning that they are actually UNDERrepresented among sick people. Cat videos might actually be correlated with good health! (Whether we have causation or not is a whole other story.) Does that help?
RiyaJ0032
Hi !
DmitryFarberMartyMurraycan you please provide an explanation for (D)
I did analyze the option but did not know how to further form an inference making (D) the ans
we are given that out of 100 people having a car accident
under 50 are (x), and above 50 are (100-x)
of which ( x/25 ) have prosinopsis
and this x/25 represent 2% of total drivers
so our total drivers are 2x
after this, how to conclude (D) to be the ans?
or if any expert could explain
Thank you so much
Frank28
Among people in Boravia who had a car accident while driving, 4 percent of those under 50 years old were found to have prosinopsis, a currently untreatable eye disease that causes a gradual deterioration In peripheral vision. Yet the proportion of drivers under 50 who have been diagnosed with prosinopsis is only 2 percent in Boravia. Clearly, therefore, having the disease makes Boravian drivers under 50 significantly more likely to have a car accident.
Which of the following is an assumption on which the argument depends?
A. Among Boravians under 50 who had a car accident while driving, most of those who had prosinopsis had the disease in its late stages.
B. Prosinopsis is not more difficult to diagnose in a person under 50 than in someone older.
C. There is no way of correcting for deterioration in peripheral vision brought about by prosinopsis.
D. Boravian drivers under 50 who have prosinopsis have generally been diagnosed as having it.
E. Prosinopsis has no symptoms other than the deterioration of peripheral vision.
Source: topgmat